I think the OP is simply trying to say that a preprint cannot be "retracted". On arxiv, it is possible to "retract" an article, in the sense that you ask that

I think the OP is simply trying to say that a preprint cannot be "retracted". On arxiv, it is possible to "retract" an article, in the sense that you ask that it's hidden or deleted etc. THat's not the same as retracting a published paper, where you usually get some justification, and a note from the editor explaining the decision to publish, and the decision to retract, and so on. More to the point, nobody cares if you "retract" a preprint, since it's expected that it may have errors not caught by peer-review, yet. No peer review means anyone can put anything they want online, and then take it back offline as they wish. Note that arxiv also gives you an option to publish different versions of a paper. So you can leave your paper with errors as a v1 and upload your post-review paper, with corrections, as v2. Again, no "retraction" needed.